Computer Chess Club Archives


Search

Terms

Messages

Subject: Proposal -- Crafty foundation code or foundation document

Author: Charles Roberson

Date: 08:11:50 02/18/05



  If Crafty were stripped of all the code Bob doesn't want copied and
  "crafty foundation" was released with Bob's restrictions. That may
  alleviate some of the issues and be a boost to the community.

   Maybe
     - take out the position evaluator and all the search code
         + the calls to the iterative deepening routine are replaced with
           a call to a random move maker.
     o the legal move generator is left in
        - maybe move ordering is taken out
        - maybe efficient routines like (popcount,leftmostbit,rightmostbit....)
           are replaced with inefficient routines
        - maybe move ordering is simplified.

    Bob could complete the specification if he likes the idea. This would
    give a simple program that plays only random moves but based on a firm
    foundation. Giving them the chore of improving some of the inefficiencies
    seems reasonable to me.

       Now, would tournaments all such a program could be a big question?

    The other possibility is the a document that describes the foundation code
    in "fair" detail. Thus a person could learn and then do the work.

     Now, an algorithmic only document could be released for the search and
    position evaluation and move ordering. It might not have much detail
    But, maybe this is telling too much.

     Also, my apologies to Bob. My inclination is to refer as Prof. Hyatt, but
   all others here refer as Bob.





This page took 0 seconds to execute

Last modified: Thu, 15 Apr 21 08:11:13 -0700

Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.